Ransomware has become one of the most prevalent and damaging forms of cyberattacks, posing a significant threat to both individuals and organizations globally. A novel hybrid detection framework that integrates Isolation Forest for anomaly detection with Long Short-Term Memory (LSTM) networks for sequence modeling offers a significant advancement in identifying ransomware activities within network traffic. The proposed approach leverages the anomaly detection capabilities of Isolation Forest to flag suspicious network flows and enhances detection accuracy through the sequential pattern recognition capabilities of LSTM, allowing it to identify ransomware communication patterns over time. Experimental results demonstrate that the combined model achieves high accuracy, precision, and recall, effectively reducing false positives and negatives while maintaining real-time processing capabilities. Through this hybrid detection method, organizations can better protect their infrastructure from ransomware attacks, ensuring rapid identification of malicious activities even as attack strategies evolve. The model's ability to operate in high-throughput network environments further underlines its relevance for large-scale cybersecurity applications.